Method of multi-dimensionally accentuating a deviation in information and identifying its cause - 6470297









The present invention is a method of accentuating a deviation in a set of information and locating a cause thereof that includes the steps of receiving a set of information; selecting a subset of information; setting initial states; setting transition states; initializing a vector; selecting the first information segment; modifying the vector if the information segment contains an initial and transition state, otherwise stopping; selecting the next available information segment and returning to the previous step, otherwise recording a number of occurrences of each unique vector; determining the number of the least occurring vector; dividing each occurrence number by the least occurring number; determining an occupation time for each vector; calculating an inverse characteristic time for each unique vector; calculating at least one subset value for the subset of information using a temperature-based function, an entropy-based function, an energy-based function, or any combination thereof; setting a value v for each initial and transition state; calculating a configuration value for each initial and transition state; selecting the subset of information then is available and next in sequence and returning to the vector modification step for further processing, otherwise plotting the subset and configuration values; finding a difference, if any, in the plotted values; and finding the information segments that correspond to the difference.Method of multi-dimensionally accentuating a deviation in information and identifying its cause647029722/10/200221/06/20002002853Ford; David K.US Patent and Trademark OfficeGoogle Patent Searchpatentimages.storage.googleapis.com/US6470297B1/US06470297-20021022-D00000.pngUniquetransition stateinformationsubsetvaluevectornumberinformation segmentunique vectorconfiguration valuesfirst information segment